LPDDR5X

Headlines · 4

1. Nov 2
   Qualcomm validates first-generation HBC silicon for Dragonfly AI250

   Qualcomm says its first-generation High Bandwidth Compute silicon completed hardware validation at its San Diego laboratory, including the full memory read-and-write path and near-memory computations. The company has taped out a data-center product using the architecture for the Dragonfly AI250 accelerator, which is expected to begin sampling in the first half of 2027. Qualcomm reports 133 TB/s of effective memory bandwidth for AI250, while its comparisons with LPDDR5X and projected HBM4 specifications are vendor-generated estimates rather than independent benchmarks.

   The Next Web · 🔥 0
2. Oct 23
   CXMT sees long-term growth in global and China DRAM

   ChangXin Technology said it remains confident in the medium- to long-term outlook for the DRAM industry in China and globally. The company said AI-related data-throughput demand continues to grow rapidly and argued that DRAM's role as strategic infrastructure will become more important, while also citing the significance of expanding China’s domestic DRAM supply.

   IT之家 AI · 🔥 0
3. Oct 12
   Intel names Crescent Island 'GPU D-Series': up to 480GB LPDDR5X for inference

   At the 2026 OCP Global Summit, Intel officially named its Crescent Island data-center inference accelerator the Intel GPU D-Series, with the first product listed as the GPU D270. According to Intel and the outlets covering it, the card uses the Xe3P architecture with 32 Xe cores and 256 XMX engines, delivering 1.3 PFLOPS of FP8 compute and 1.5TB/s of memory bandwidth at a 350W TDP, in a standard PCIe Gen5, air-cooled form factor. Intel says one card can run models of roughly 300 billion parameters and an eight-GPU system can handle over one trillion parameters, though 480GB is the ceiling opened to ODM partners while Intel-branded PCIe cards carry 160GB. Intel says sampling is underway; pricing and a general-availability date have not been announced.

   Tom’s Hardware · 🔥 0
4. Oct 8
   Samsung may cut Q4 smartphone output by up to 30% as AI-driven memory prices bite

   Multiple IT and Korean media reports say Samsung's Mobile Experience business has asked component suppliers to trim shipments by 20%–30%, with fourth-quarter smartphone production possibly falling by as much as 30% because soaring memory prices, driven by AI demand, leave little profit in selling phones. The reports cite market data showing mobile LPDDR5X memory at roughly $145–146 per 12GB in Q2, up 175% year on year, with a further ~20% rise possible in Q3 toward $180; Samsung's first-half phone memory procurement costs reportedly jumped 211% on average while average selling prices rose only about 7%. Samsung had estimated 2026 smartphone output of 270 million units, but a Q4 cut could pull full-year shipments down to 220–230 million.

   科技新报 · 🔥 0

Experience and discussion from the community

Share my LPDDR5X experienceAsk about LPDDR5X